### Table 3: Complexity and relative computation time of block-based methods Method and reference(s) Complexity Time

### Table 3. Dot product cycle count analysis of basic blocks based on n

"... In PAGE 89: ... 3. Prequential error for EDDM and DDM in ELEC2 dataset Table3 . Prequential error and total number of changes detected by methods in ELEC2 dataset EDDM DDM Prequential Drifts Prequential Drifts IB1 0.... In PAGE 89: ...2110 130 can be seen, EDDM outperforms DDM, detecting concept changes earlier and with a better sensitivity. Table3 shows the flnal prequential errors and number of drifts obtained by these two methods on the electricity market dataset. 5 Conclusion This paper introduces a new method for detecting concept drifts based on the distances between classiflcation errors.... In PAGE 105: ... kernel. A result of this analysis is presented in Table3 . Thus, the mini- mum cycle count for n = 1 is 261 cycles (assuming that the kernelDot method is the cache.... In PAGE 106: ... Accordingly, it is then possible to construct the table for the SVM decision function, which then depicts the formula for cycle count based on m, n, and sv. This time (as opposed to Table3 ) the min. cycle count is omitted.... In PAGE 147: ...Where, Vf and Ef are number of Vertices and Edges in the compressed graph and Vi and Ei are the respective numbers in the original graphs. Table3 . Results for compression Quotient Sample Number of nodes Q1 Q2 Q3 Sample1 100K 0.... ..."

### Table 3. Dot product cycle count analysis of basic blocks based on n

"... In PAGE 5: ... kernel. A result of this analysis is presented in Table3 . Thus, the mini- mum cycle count for n = 1 is 261 cycles (assuming that the kernelDot method is the cache.... In PAGE 6: ... Accordingly, it is then possible to construct the table for the SVM decision function, which then depicts the formula for cycle count based on m, n, and sv. This time (as opposed to Table3 ) the min. cycle count is omitted.... ..."

### Table 2 summarizes the performance of the proposed, block- based, and average methods [5]. As shown, the proposed method gives significantly lower error than the reference methods. Table 2) also show that the proposed method is four times faster than the block-based method which has been found to be the most com- putationally efficient among tested noise estimation methods [5].

2002

Cited by 5

### Table 2: Comparison of PVS sizes for block-based (125 blocks) and building-based (665 buildings) PVS computation. Going from block-based to building-based PVSs reduced the average difference between the two methods from 10.91% to 2.98% of the complete model.

2000

"... In PAGE 7: ... This supports the need for methods such as the one presented in this paper. We give the comparison between our PVS and the point-sampled solution in Table2 . Note that the difference is always positive or zero, which demonstrates that our method is conservative.... ..."

Cited by 49

### Table 2: Comparison of PVS sizes for block-based (125 blocks) and building-based (665 buildings) PVS computation. Going from block-based to building-based PVSs reduced the average difference between the two methods from 10.91% to 2.98% of the complete model.

"... In PAGE 7: ... This supports the need for methods such as the one presented in this paper. We give the comparison between our PVS and the point-sampled solution in Table2 . Note that the difference is always positive or zero, which demonstrates that our method is conservative.... ..."

### Table 3. Performance of block based JAD MVP algorithm.

"... In PAGE 8: ...Block based algorithms perform much better than the point based ones. For small test cases the performance of MVP for block sizes 4 and 5 is listed in Table3 . Operating on small dense blocks is a superior way to achieve a good percentage of the peak performance on vector machines (37% for block size 5).... ..."

### Table 2: Comparison of running time between the optimized algorithm and the block-based algorithm.

2005

"... In PAGE 8: ... EXPERIENCE WITH ATOMICITY CHECK- ING We evaluated the focused block-based atomicity checking al- gorithm on the flrst six benchmarks described in the previ- ous section. The results appear in Table2 . Running times are measured in seconds on a 1GHz Sun Blade 1500 with Sun JDK 1.... ..."

Cited by 13

### Table 2: Comparison of running time between the optimized algorithm and the block-based algorithm.

2005

"... In PAGE 8: ... EXPERIENCE WITH ATOMICITY CHECKING We evaluated the focused block-based atomicity checking algorithm on the benchmarks described in the previous sec- tion except Jigsaw, because Soot failed to construct the call graph for it. The results appear in Table2 . Running times are measured in seconds on a 1GHz Sun Blade 1500 with Sun JDK 1.... ..."

Cited by 13

### Table 5.4: Average spectral distortion and 2 dB outliers percentage for Block-based

1999

Cited by 1